The present invention relates to computing systems, and more specifically, this invention relates to systems, methods and computer program products for implementing an automated shutdown of a tiered system.
Conventional computing systems, for example, enterprise applications, typically possess tiered architectures. Tiered systems are typically constructed from a collection of components that combine to implement high-performance computing applications. Each of the components in the tiered system may be a separate, independent computing device interconnected with the other components via a communication medium, such as a network.
Tiered systems provide specialized solutions catering to different business aspects within an organization or across geographically distant installations. These systems further enable enterprises to operate critical tasks (e.g. data storage, archiving business records, etc.) with reduced involvement and individual know-how of the end-users (e.g. customers).
Management of tiered computing systems to maintain architectural integrity and performance of the computing systems is critical for providing continuous business operations. In the course of managing a tiered computing system, a situation may arise that requires the shutdown of a component and/or the entire system. For example, a potential or actual emergency (e.g., an earthquake, flood, fire, etc) or other immediate crises (e.g., a power outage) may warrant shutdown. Additionally, shutdown may be necessary when an individual component and/or node within the multi-tiered system reaches a thermal failure or is exposed to an unexpected environment (e.g., blocked vents, fan failures, lab cooling failures, etc.).
Typically, the shutdown of a tiered system is complex and must occur according to a specified sequence. Improper shutdown may have disastrous effects, some as severe as the loss of customer data and loss of configuration. Consequently, the shutdown of a tiered system is usually only performed by a certified human operator and thus may be limited to the human operator's response time.